US20220022816A1 - Method and system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome - Google Patents
Method and system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome Download PDFInfo
- Publication number
- US20220022816A1 US20220022816A1 US17/311,718 US202017311718A US2022022816A1 US 20220022816 A1 US20220022816 A1 US 20220022816A1 US 202017311718 A US202017311718 A US 202017311718A US 2022022816 A1 US2022022816 A1 US 2022022816A1
- Authority
- US
- United States
- Prior art keywords
- exosome
- data
- blood
- deep learning
- sers
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7271—Specific aspects of physiological measurement analysis
- A61B5/7275—Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/65—Raman scattering
- G01N21/658—Raman scattering enhancement Raman, e.g. surface plasmons
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/145—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue
- A61B5/1455—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue using optical sensors, e.g. spectral photometrical oximeters
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
- G16B40/10—Signal processing, e.g. from mass spectrometry [MS] or from PCR
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
- G16B40/20—Supervised data analysis
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
Definitions
- the present disclosure relates to a method and system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome.
- Exosomes are extracellular vesicles that are rich in body fluids such as blood, urine and saliva, and are materials that are secreted from all cells and involved in a variety of biological mechanisms such as interaction between cells and metastasis.
- the exosomes are 30-200 nm in size and produced by forming multivesicular bodies intracellularly by endocytosis, which are released by exocytosis, and include materials representative of the characteristics of parent cells such as protein and miRNA. Accordingly, the exosomes containing the characteristics of disease cells are present in body fluids, and may be used as biomarkers for cancer diagnosis based on these characteristics.
- Raman scattering when a molecule is illuminated with light, inelastic scattering that may exhibit the characteristics of vibrational modes of bonds occurs in the molecule, and it is called Raman scattering. Since the vibrational modes of the molecule appear according to the characteristics of bonds in the molecule, a Raman scattered light signal found in the molecule may be used as fingerprint information of the molecule.
- signal sensitivity that is so low that the frequency of Raman scattering is one in every million photons.
- SERS Surface Enhanced Raman Spectroscopy
- SERS is a technique that can acquire Raman spectroscopy signals found on the surface of exosomes with ultrahigh sensitivity, and thus is effective in detecting the characteristics of exosomes specific to diseases. Additionally, since SERS is a label-free detection technique, a preprocessing process for capturing only a targeted marker may be omitted and high signal sensitivity is achieved even in a low concentration sample.
- the SERS signals of exosomes detect heterogeneous signals and are complex signals due to heterogeneous materials between individual exosomes, interaction with plasmonic nanoparticles which are probes and material distributions of attached surfaces.
- PCA principal component analysis
- PLS-DA partial least-squares discriminant analysis
- an embodiment of the present disclosure provides a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome.
- the method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome includes measuring a cultured cell exosome Surface Enhanced Raman Spectroscopy (SERS) signal, training a deep learning model through the cultured cell exosome SERS signal, measuring a blood exosome SERS signal, analyzing the blood exosome SERS signal through the deep learning model trained with the cultured cell exosome SERS signal, and analyzing a similarity between blood exosome data and cultured cell exosome data analyzed through the deep learning model.
- SERS Surface Enhanced Raman Spectroscopy
- another embodiment of the present disclosure provides a system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome.
- the system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome includes a training data input unit to receive input of a cultured cell exosome SERS signal used to train a deep learning model, a patient data input unit to receive input of a patient's blood exosome SERS signal data, the deep learning model trained using the cultured cell exosome SERS signal inputted through the training data input unit to analyze the blood exosome SERS signal data inputted through the patient data input unit, and a similarity analysis unit to analyze a similarity between blood exosome data and cultured cell exosome data analyzed through the deep learning model.
- SERS Surface Enhanced Raman Spectroscopy
- FIG. 1 is a flowchart of a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure.
- FIG. 2 is a diagram for describing a deep learning model according to an embodiment of the present disclosure.
- FIG. 3 is a diagram for describing a data analysis step according to an embodiment of the present disclosure.
- FIG. 4 is a diagram for describing a similarity analysis step according to an embodiment of the present disclosure.
- FIG. 5 is a diagram showing normal cell and lung cancer cell exosome Raman signals acquired according to an embodiment of the present disclosure.
- FIG. 6 is a diagram showing a result of classifying normal cell and lung cancer cell exosome Raman signals through a deep learning model according to an embodiment of the present disclosure.
- FIG. 7 is a diagram showing accuracy after data training according to an embodiment of the present disclosure.
- FIG. 8 is a diagram showing blood exosome Raman signals of a normal person and a lung cancer patient acquired according to an embodiment of the present disclosure.
- FIG. 9 is a diagram showing a result of analyzing blood exosome Raman signals of a normal person and a lung cancer patient according to an embodiment of the present disclosure.
- FIG. 10 is a diagram showing a receiver operating characteristic (ROC) curve for a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure.
- ROC receiver operating characteristic
- FIG. 11 is a diagram showing sensitivity and specificity by a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure.
- FIG. 12 is a diagram showing a result of staging lung cancer using a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure.
- FIG. 13 is a block diagram of a system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to another embodiment of the present disclosure.
- FIG. 1 is a flowchart of a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure.
- the method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome may include the steps of (S 110 ) measuring cultured cell exosome Surface Enhanced Raman Spectroscopy (SERS) signals, (S 120 ) training a deep learning model through the cultured cell exosome SERS signals, (S 130 ) measuring blood exosome SERS signals, (S 140 ) analyzing the blood exosome SERS signals through the deep learning model trained with the cultured cell exosome SERS signals, and (S 150 ) analyzing a similarity between the blood exosome data and cultured cell exosome data analyzed through the deep learning model.
- SERS Surface Enhanced Raman Spectroscopy
- the cultured cell exosome SERS signals may be measured.
- the cultured cells may include cultured normal cells and cancer cells.
- an agglomerated structure of 80 nm spherical gold nanoparticles using salts may be used to measure the SERS signals of exosomes.
- the corresponding structure may be formed by spraying concentrated gold nanoparticles onto a washed cover glass, adding a CuSO 4 based salt solution to cause gold nanoparticles to agglomerate, and drying.
- the SERS signals may be measured after applying an exosome solution with a concentration of 10 7 to 10 9 particles/mL to the manufactured substrate and completely drying.
- the deep learning model may be trained using the measured cultured cell exosome SERS signals, i.e., SERS signals of normal cell exosomes and SERS signals of cancer cell exosomes derived from cultured normal cells and cancer cells, respectively.
- an applied model of ResNet model may be used as the deep learning model.
- FIG. 2 is a diagram for describing the deep learning model according to an embodiment of the present disclosure
- (A) of FIG. 2 shows the ResNet model
- (B) shows the applied model of the ResNet model according to an embodiment of the present disclosure.
- the ResNet model shown in (A) of FIG. 2 was proposed in 2015 (He, Kaiming, et al. “Deep residual learning for image recognition.” Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.), and basically, was developed for 2-dimensional image analysis.
- the ResNet model shown in (A) of FIG. 2 is adapted as shown in (B) of FIG. 2 and used to analyze SERS signal data which is 1-dimensional data.
- the ResNet model shown in (A) of FIG. 2 is a 34-layer ResNet model and includes a first convolutional layer cony. 64, /2 where data is inputted, a plurality of second convolutional layers cony. 64, a plurality of third convolutional layers cony. 128, a plurality of fourth convolutional layers cony. 256 and a plurality of fifth convolutional layers cony. 512 connected to the first convolutional layer in that order, and a fully connected layer (FCL), and the second to fifth convolutional layers are configured to perform residual learning by adding skip connection every two convolutional layers.
- FCL fully connected layer
- the deep learning model according to an embodiment of the present disclosure as shown in (B) of FIG. 2 has a simple convolutional layer configuration, and includes a first convolutional layer cony. 64 where data is inputted, 4 second convolutional layers cony. 64, 4 third convolutional layers cony. 128, 4 fourth convolutional layers cony. 256 and 4 fifth convolutional layers cony. 512 connected to the first convolutional layer in that order, and 2 fully connected layers FCL 4096, FCL 10, and the second to fifth convolutional layers add skip connection every two convolutional layers.
- the number of convolutional layers included in the second to fifth convolutional layers is not limited thereto and may change where necessary.
- the fully connected layer (FCL) located at the end of the deep learning model includes a first fully connected layer FCL 4096 having 4096 values, and a second fully connected layer FCL 10 connected to the output end of the first fully connected layer and having 10 values, and finally, may classify data into normal exosome and lung cancer exosome.
- the following embodiment preprocesses SERS signal data through, for example, min-max normalization before training the deep learning model, and uses Raman spectrum in the range of 475 to 2000 cm ⁇ 1 as data.
- an embodiment of the present disclosure may train the deep learning model using SERS signals of exosomes in cultured cells, i.e., normal cells and cancer cells, thereby classifying the characteristics of the normal cells and the cancer cells more accurately.
- SERS signals of exosomes in the patient's blood may be measured.
- blood exosome SERS signals may be measured by the same process as described above in the step of measuring the cultured cell exosome SERS signals (S 110 ), and its repetitive description is omitted herein.
- blood exosome SERS signals may be analyzed using the deep learning model trained in the above-described step S 120 .
- exosomes in the patient's blood include normal cell derived exosomes and cancer cell derived exosomes together, it is difficult to classify them by binary classification like as classifying cultured cell exosomes.
- FIG. 3 is a diagram for describing the data analysis step according to an embodiment of the present disclosure.
- the FCL 10 has 10 numerical values as values derived through all the previous layers, and when the blood exosome SERS signals are classified using the trained deep learning model, 10 values may be derived from FCL 10.
- analysis may be conducted to determine whether exosome data in the patient's blood is similar to normal cell exosome data or cancer cell exosome data, by comparing FCL 10 values derived by classifying the blood exosome SERS signals with FCL 10 values derived by classifying SERS signals of exosomes in cultured cells, i.e., normal cells and cancer cells respectively.
- FIG. 4 is a diagram for describing the similarity analysis step according to an embodiment of the present disclosure, and for convenience, showing 10 values derived from the FCL 10 on 2-dimensions.
- the similarity between blood exosome data and cultured cell exosome data may be calculated through Mahalanobis distance.
- the similarity calculation method according to the present disclosure is not limited thereto, and a variety of methods for quantitating similarities between data may be applied.
- the calculated Mahalanobis distances may indicate de-similarity between each data.
- a method of calculating the Mahalanobis distance is well known to those skilled in the art, and its detailed description is omitted herein.
- a relative similarity to cancer exosome i.e., a relative similarity of blood exosome data to cancer cell exosome data may be calculated by the following equation.
- Relative similarity to cancer exosome Mahalanobis distance D normal of blood exosome data from normal cell exosome data/Mahalanobis distance D cancer of blood exosome data from cancer cell exosome data
- the relative similarity of blood exosome data to cancer cell exosome data derived through the above-described process may be used to diagnose cancers.
- the present disclosure is described based on an embodiment applied to lung cancer patients.
- the present disclosure is not necessarily limited thereto, and in addition to lung cancer, may be equally to patients with, for example, pancreatic cancer, breast cancer and blood cancer.
- FIG. 5 is a diagram showing normal cell and lung cancer cell exosome Raman signals acquired according to an embodiment of the present disclosure, showing 340 SERS signals of exosomes derived from each of normal lung cells (human pulmonary alveolar epithelial cells (HPAEC)) and 3 types of lung cancer cells (PC9, H522, H1299).
- normal lung cells human pulmonary alveolar epithelial cells (HPAEC)
- HPAEC human pulmonary alveolar epithelial cells
- FIG. 6 is a diagram showing a result of classifying normal cell and lung cancer cell exosome Raman signals through the deep learning model according to an embodiment of the present disclosure.
- normal cell and lung cancer cell exosome Raman signals are classified through the deep learning model according to an embodiment of the present disclosure, it can be seen that normal exosomes and lung cancer exosomes are classified as shown in FIG. 6 .
- Table 1 shows a result of validating the deep learning model according to an embodiment of the present disclosure through 10-CV (10 cross validation).
- the validation results reveal that the deep learning model according to an embodiment of the present disclosure shows 96% average classification accuracy for trained exosomes, and 80% average classification accuracy for untrained exosome data.
- FIG. 7 is a diagram showing accuracy after data training according to an embodiment of the present disclosure, showing 98% classification accuracy when all data is used in training.
- FIG. 8 is a diagram showing blood exosome Raman signals of normal persons and lung cancer patients acquired according to an embodiment of the present disclosure, showing the mean of exosome SERS signal data acquired from 24 normal persons (healthy control) and 18 lung cancer patients (NSCLS patients).
- the similarity to cultured cell exosome data may be analyzed by analysis through the deep learning model for each individual data as described above.
- FIG. 9 is a diagram showing a result of analyzing the blood exosome Raman signals of normal persons and lung cancer patients according to an embodiment of the present disclosure.
- FIG. 9 show a heatmap displaying the analysis results of 50 exosome data for each person of 24 normal persons and 18 lung cancer patients respectively, showing the closer to red, the higher the relative similarity to lung cancer cell exosomes.
- (C) of FIG. 9 statistically shows the mean of 50 exosome data acquired from one person. A higher similarity to lung cancer cell exosomes is found in lung cancer patients than normal persons, and a p-value is 0.000036, and thus it can be seen that a statistical difference is very significant.
- FIG. 10 is a diagram showing a receiver operating characteristic (ROC) curve for a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure.
- ROC receiver operating characteristic
- the ROC curve is shown to validate the effect of the deep learning model according to the above-described embodiment of the present disclosure.
- area under the ROC curve i.e., area under curve (AUC) is closer to 1, utility is higher, and the deep learning model according to an embodiment of the present disclosure shows a value of 0.89.
- Table 2 shows AUC value comparison results of the deep learning model according to an embodiment of the present disclosure and other research findings.
- FIG. 11 is a diagram showing sensitivity and specificity by the method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure.
- a diagnostic cutoff point (cutoff line) is set to 0.85 based on the value derived from the ROC curve, sensitivity and specificity are about 89% and 88% respectively. That is, it can be seen that it is possible to clearly distinguish normal persons and lung cancer patients according to an embodiment of the present disclosure.
- FIG. 12 is a diagram showing a result of staging lung cancer using the method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure, and in the graph, the stage is higher as it goes rightward.
- the above-described embodiment of the present disclosure may be used to stage cancers.
- it may be used in a manner of setting a similarity reference value by grade for each cancer stage and providing cancer stage information according to a range corresponding to the similarity derived by the process described above with reference to FIG. 1 .
- FIG. 13 is a block diagram of a system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to another embodiment of the present disclosure.
- the system 1000 for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome may include a training data input unit 1100 , a patient data input unit 1200 , a deep learning model 1300 and a similarity analysis unit 1400 .
- the training data input unit 1100 may receive input of exosome SERS signal data used to train the deep learning model 1300 .
- the training data input unit 1100 may receive input of SERS signals of exosomes derived from each of cultured cells, i.e., normal cells and cancer cells.
- the patient data input unit 1200 may receive input of blood exosome SERS signals of a patient for providing cancer diagnosis information.
- the deep learning model 1300 may first perform learning using the normal cell and cancer cell exosome SERS signals inputted through the training data input unit 1100 .
- the deep learning model 1300 may analyze the blood exosome SERS signals inputted through the patient data input unit 1200 .
- the similarity analysis unit 1400 may analyze a similarity between the blood exosome data and cultured cell exosome data analyzed through the deep learning model 1300 .
- the system for providing cancer diagnosis information shown in FIG. 13 may be implemented in a computing device capable of deep learning computation.
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Medical Informatics (AREA)
- Biophysics (AREA)
- Public Health (AREA)
- Pathology (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Veterinary Medicine (AREA)
- Heart & Thoracic Surgery (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Epidemiology (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Databases & Information Systems (AREA)
- Psychiatry (AREA)
- Physiology (AREA)
- Immunology (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Fuzzy Systems (AREA)
- Primary Health Care (AREA)
- Evolutionary Biology (AREA)
Abstract
A method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure includes measuring a cultured cell exosome Surface Enhanced Raman Spectroscopy (SERS) signal, training a deep learning model through the cultured cell exosome SERS signal, measuring a blood exosome SERS signal, analyzing the blood exosome SERS signal through the deep learning model trained with the cultured cell exosome SERS signal, and analyzing a similarity between blood exosome data and cultured cell exosome data analyzed through the deep learning model.
Description
- The present disclosure relates to a method and system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome.
- Exosomes are extracellular vesicles that are rich in body fluids such as blood, urine and saliva, and are materials that are secreted from all cells and involved in a variety of biological mechanisms such as interaction between cells and metastasis. The exosomes are 30-200 nm in size and produced by forming multivesicular bodies intracellularly by endocytosis, which are released by exocytosis, and include materials representative of the characteristics of parent cells such as protein and miRNA. Accordingly, the exosomes containing the characteristics of disease cells are present in body fluids, and may be used as biomarkers for cancer diagnosis based on these characteristics.
- Meanwhile, when a molecule is illuminated with light, inelastic scattering that may exhibit the characteristics of vibrational modes of bonds occurs in the molecule, and it is called Raman scattering. Since the vibrational modes of the molecule appear according to the characteristics of bonds in the molecule, a Raman scattered light signal found in the molecule may be used as fingerprint information of the molecule. However, its disadvantage is signal sensitivity that is so low that the frequency of Raman scattering is one in every million photons.
- Surface Enhanced Raman Spectroscopy (SERS) is a method designed to solve the low signal intensity problem of normal Raman Spectroscopy, and may amplify a Raman signal of the molecule about 107-108 times or more based on a strong electromagnetic field in a nanogap between plasmonic nanostructures. Technically, it is said that this method achieves single molecule-level detection at the femtomolar concentration levels of molecules.
- As described above, since the exosomes contain a lot of information of parent cells, attention is focused on the use as biomarkers for liquid biopsy in the diagnostic medical field. Studies have been made to detect information of disease exosomes in body fluids and early diagnose or monitor the diseases by applying a variety of biological material detection technologies. In particular, it is predicted to solve the problems of the existing cancer diagnosis techniques such as invasive diagnosis, high false positive risks and exposure to very high levels of radiation through the exosome based liquid biopsy technique.
- SERS is a technique that can acquire Raman spectroscopy signals found on the surface of exosomes with ultrahigh sensitivity, and thus is effective in detecting the characteristics of exosomes specific to diseases. Additionally, since SERS is a label-free detection technique, a preprocessing process for capturing only a targeted marker may be omitted and high signal sensitivity is achieved even in a low concentration sample.
- However, the SERS signals of exosomes detect heterogeneous signals and are complex signals due to heterogeneous materials between individual exosomes, interaction with plasmonic nanoparticles which are probes and material distributions of attached surfaces.
- To overcome these limitations, methods such as principal component analysis (PCA) and partial least-squares discriminant analysis (PLS-DA) have been applied, but it is difficult to analyze cancer exosomes in patients' body fluids since they are present together with exosomes derived from normal cells.
- There is a need for an approach to provide cancer diagnosis information more accurately using Surface Enhanced Raman Spectroscopy (SERS) signals of exosomes in the technical field.
- To solve the above-described technical problem, an embodiment of the present disclosure provides a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome.
- The method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome includes measuring a cultured cell exosome Surface Enhanced Raman Spectroscopy (SERS) signal, training a deep learning model through the cultured cell exosome SERS signal, measuring a blood exosome SERS signal, analyzing the blood exosome SERS signal through the deep learning model trained with the cultured cell exosome SERS signal, and analyzing a similarity between blood exosome data and cultured cell exosome data analyzed through the deep learning model.
- In addition, another embodiment of the present disclosure provides a system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome.
- The system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome includes a training data input unit to receive input of a cultured cell exosome SERS signal used to train a deep learning model, a patient data input unit to receive input of a patient's blood exosome SERS signal data, the deep learning model trained using the cultured cell exosome SERS signal inputted through the training data input unit to analyze the blood exosome SERS signal data inputted through the patient data input unit, and a similarity analysis unit to analyze a similarity between blood exosome data and cultured cell exosome data analyzed through the deep learning model.
- Furthermore, the above-described technical solution does not enumerate all the features of the present disclosure. A variety of features of the present disclosure and their advantages and effects will be understood in more detail with reference to the following specific embodiments.
- According to an embodiment of the present disclosure, it is possible to overcome the limitation of signal heterogeneity through analysis of complex Surface Enhanced Raman Spectroscopy (SERS) signals of exosomes via artificial intelligence. Through this, it is possible to achieve noninvasive cancer liquid biopsy using exosomes. Additionally, it is possible to diagnose cancer using only blood without risk of exposure to radiation such as X-ray and Computer Tomography (CT) or invasive tissue biopsy, and in addition to cancer diagnosis, to monitor patients' treatment.
-
FIG. 1 is a flowchart of a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure. -
FIG. 2 is a diagram for describing a deep learning model according to an embodiment of the present disclosure. -
FIG. 3 is a diagram for describing a data analysis step according to an embodiment of the present disclosure. -
FIG. 4 is a diagram for describing a similarity analysis step according to an embodiment of the present disclosure. -
FIG. 5 is a diagram showing normal cell and lung cancer cell exosome Raman signals acquired according to an embodiment of the present disclosure. -
FIG. 6 is a diagram showing a result of classifying normal cell and lung cancer cell exosome Raman signals through a deep learning model according to an embodiment of the present disclosure. -
FIG. 7 is a diagram showing accuracy after data training according to an embodiment of the present disclosure. -
FIG. 8 is a diagram showing blood exosome Raman signals of a normal person and a lung cancer patient acquired according to an embodiment of the present disclosure. -
FIG. 9 is a diagram showing a result of analyzing blood exosome Raman signals of a normal person and a lung cancer patient according to an embodiment of the present disclosure. -
FIG. 10 is a diagram showing a receiver operating characteristic (ROC) curve for a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure. -
FIG. 11 is a diagram showing sensitivity and specificity by a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure. -
FIG. 12 is a diagram showing a result of staging lung cancer using a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure. -
FIG. 13 is a block diagram of a system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to another embodiment of the present disclosure. - Hereinafter, the preferred embodiments will be described in sufficiently detail for those skilled in the art to easily practice the invention with reference to the accompanying drawings. However, in describing the preferred embodiments of the present disclosure in detail, when a certain detailed description of relevant known functions or elements is deemed as rendering the subject matter of the present disclosure unnecessarily ambiguous, the detailed description is omitted herein. Additionally, elements with similar functions and operations use the same signs throughout the drawings.
- Additionally, in the specification, when an element is referred to as being ‘connected to’ another element, the element may be ‘directly connected to’ the other element and the element may be ‘indirectly connected to’ the other element with intervening elements between them. Additionally, the term ‘comprises’ when used in this specification, specifies the presence of stated elements, does not preclude the presence or addition of one or more other elements unless the context clearly indicates otherwise.
-
FIG. 1 is a flowchart of a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure. - Referring to
FIG. 1 , the method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure may include the steps of (S110) measuring cultured cell exosome Surface Enhanced Raman Spectroscopy (SERS) signals, (S120) training a deep learning model through the cultured cell exosome SERS signals, (S130) measuring blood exosome SERS signals, (S140) analyzing the blood exosome SERS signals through the deep learning model trained with the cultured cell exosome SERS signals, and (S150) analyzing a similarity between the blood exosome data and cultured cell exosome data analyzed through the deep learning model. - First, in the step of measuring the cultured cell exosome SERS signals (S110), SERS signals of exosomes derived from cultured cells may be measured. Here, the cultured cells may include cultured normal cells and cancer cells.
- According to an embodiment, an agglomerated structure of 80 nm spherical gold nanoparticles using salts may be used to measure the SERS signals of exosomes. Here, the corresponding structure may be formed by spraying concentrated gold nanoparticles onto a washed cover glass, adding a CuSO4 based salt solution to cause gold nanoparticles to agglomerate, and drying. The SERS signals may be measured after applying an exosome solution with a concentration of 107 to 109 particles/mL to the manufactured substrate and completely drying.
- In the step of training the deep learning model through the cultured cell exosome SERS signals (S120), the deep learning model may be trained using the measured cultured cell exosome SERS signals, i.e., SERS signals of normal cell exosomes and SERS signals of cancer cell exosomes derived from cultured normal cells and cancer cells, respectively.
- According to an embodiment, an applied model of ResNet model may be used as the deep learning model.
-
FIG. 2 is a diagram for describing the deep learning model according to an embodiment of the present disclosure, (A) ofFIG. 2 shows the ResNet model, and (B) shows the applied model of the ResNet model according to an embodiment of the present disclosure. - The ResNet model shown in (A) of
FIG. 2 was proposed in 2015 (He, Kaiming, et al. “Deep residual learning for image recognition.” Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.), and basically, was developed for 2-dimensional image analysis. - According to an embodiment of the present disclosure, the ResNet model shown in (A) of
FIG. 2 is adapted as shown in (B) ofFIG. 2 and used to analyze SERS signal data which is 1-dimensional data. - Describing in detail, the ResNet model shown in (A) of
FIG. 2 is a 34-layer ResNet model and includes a first convolutional layer cony. 64, /2 where data is inputted, a plurality of second convolutional layers cony. 64, a plurality of third convolutional layers cony. 128, a plurality of fourth convolutional layers cony. 256 and a plurality of fifth convolutional layers cony. 512 connected to the first convolutional layer in that order, and a fully connected layer (FCL), and the second to fifth convolutional layers are configured to perform residual learning by adding skip connection every two convolutional layers. - In contrast, the deep learning model according to an embodiment of the present disclosure as shown in (B) of
FIG. 2 has a simple convolutional layer configuration, and includes a first convolutional layer cony. 64 where data is inputted, 4 second convolutional layers cony. 64, 4 third convolutional layers cony. 128, 4 fourth convolutional layers cony. 256 and 4 fifth convolutional layers cony. 512 connected to the first convolutional layer in that order, and 2 fully connectedlayers FCL 4096,FCL 10, and the second to fifth convolutional layers add skip connection every two convolutional layers. Here, the number of convolutional layers included in the second to fifth convolutional layers is not limited thereto and may change where necessary. Additionally, the fully connected layer (FCL) located at the end of the deep learning model includes a first fully connectedlayer FCL 4096 having 4096 values, and a second fully connectedlayer FCL 10 connected to the output end of the first fully connected layer and having 10 values, and finally, may classify data into normal exosome and lung cancer exosome. - The following embodiment preprocesses SERS signal data through, for example, min-max normalization before training the deep learning model, and uses Raman spectrum in the range of 475 to 2000 cm−1 as data.
- As described above, an embodiment of the present disclosure may train the deep learning model using SERS signals of exosomes in cultured cells, i.e., normal cells and cancer cells, thereby classifying the characteristics of the normal cells and the cancer cells more accurately.
- In the step of measuring the blood exosome SERS signals (S130), SERS signals of exosomes in the patient's blood may be measured. In this case, blood exosome SERS signals may be measured by the same process as described above in the step of measuring the cultured cell exosome SERS signals (S110), and its repetitive description is omitted herein.
- In the step of analyzing the blood exosome SERS signals through the deep learning model trained with the cultured cell exosome SERS signals (S140), blood exosome SERS signals may be analyzed using the deep learning model trained in the above-described step S120.
- Since the exosomes in the patient's blood include normal cell derived exosomes and cancer cell derived exosomes together, it is difficult to classify them by binary classification like as classifying cultured cell exosomes.
- Accordingly, according to an embodiment of the present disclosure, it proposes making use of information of the second fully connected
layer FCL 10 that is an immediately previous layer, not the final binary output layer, as described below.FIG. 3 is a diagram for describing the data analysis step according to an embodiment of the present disclosure. - As shown in
FIG. 3 , theFCL 10 has 10 numerical values as values derived through all the previous layers, and when the blood exosome SERS signals are classified using the trained deep learning model, 10 values may be derived fromFCL 10. - In the step of analyzing the similarity between the analyzed blood exosome data and the cultured cell exosome data (S150), analysis may be conducted to determine whether exosome data in the patient's blood is similar to normal cell exosome data or cancer cell exosome data, by comparing
FCL 10 values derived by classifying the blood exosome SERS signals withFCL 10 values derived by classifying SERS signals of exosomes in cultured cells, i.e., normal cells and cancer cells respectively. -
FIG. 4 is a diagram for describing the similarity analysis step according to an embodiment of the present disclosure, and for convenience, showing 10 values derived from theFCL 10 on 2-dimensions. - According to an embodiment, the similarity between blood exosome data and cultured cell exosome data may be calculated through Mahalanobis distance. However, the similarity calculation method according to the present disclosure is not limited thereto, and a variety of methods for quantitating similarities between data may be applied.
- In detail, in the picture shown on the left side of
FIG. 4 , when Mahalanobis distances between blood exosome data (Human exosome data) and normal cell exosome data (Mean of normal exosomes), and between blood exosome data (Human exosome data) and cancer cell exosome data (Mean of cancer exosomes) are calculated, the calculated Mahalanobis distances may indicate de-similarity between each data. Here, a method of calculating the Mahalanobis distance is well known to those skilled in the art, and its detailed description is omitted herein. - According to an embodiment of the present disclosure, a relative similarity to cancer exosome, i.e., a relative similarity of blood exosome data to cancer cell exosome data may be calculated by the following equation.
- Relative similarity to cancer exosome=Mahalanobis distance Dnormal of blood exosome data from normal cell exosome data/Mahalanobis distance Dcancer of blood exosome data from cancer cell exosome data
- Accordingly, according to an embodiment of the present disclosure, the relative similarity of blood exosome data to cancer cell exosome data derived through the above-described process may be used to diagnose cancers.
- Hereinafter, the present disclosure is described based on an embodiment applied to lung cancer patients. However, the present disclosure is not necessarily limited thereto, and in addition to lung cancer, may be equally to patients with, for example, pancreatic cancer, breast cancer and blood cancer.
-
FIG. 5 is a diagram showing normal cell and lung cancer cell exosome Raman signals acquired according to an embodiment of the present disclosure, showing 340 SERS signals of exosomes derived from each of normal lung cells (human pulmonary alveolar epithelial cells (HPAEC)) and 3 types of lung cancer cells (PC9, H522, H1299). -
FIG. 6 is a diagram showing a result of classifying normal cell and lung cancer cell exosome Raman signals through the deep learning model according to an embodiment of the present disclosure. - When normal cell and lung cancer cell exosome Raman signals are classified through the deep learning model according to an embodiment of the present disclosure, it can be seen that normal exosomes and lung cancer exosomes are classified as shown in
FIG. 6 . - The following Table 1 shows a result of validating the deep learning model according to an embodiment of the present disclosure through 10-CV (10 cross validation).
-
TABLE 1 CV Train accuracy(%) Test accuracy(%) 1 95 64 2 98 83 3 98 85 4 95 88 5 95 83 6 96 77 7 98 51 8 96 70 9 95 94 10 95 91 Mean 96 80 - The validation results reveal that the deep learning model according to an embodiment of the present disclosure shows 96% average classification accuracy for trained exosomes, and 80% average classification accuracy for untrained exosome data.
-
FIG. 7 is a diagram showing accuracy after data training according to an embodiment of the present disclosure, showing 98% classification accuracy when all data is used in training. -
FIG. 8 is a diagram showing blood exosome Raman signals of normal persons and lung cancer patients acquired according to an embodiment of the present disclosure, showing the mean of exosome SERS signal data acquired from 24 normal persons (healthy control) and 18 lung cancer patients (NSCLS patients). - The similarity to cultured cell exosome data may be analyzed by analysis through the deep learning model for each individual data as described above.
-
FIG. 9 is a diagram showing a result of analyzing the blood exosome Raman signals of normal persons and lung cancer patients according to an embodiment of the present disclosure. - (A) and (B) of
FIG. 9 show a heatmap displaying the analysis results of 50 exosome data for each person of 24 normal persons and 18 lung cancer patients respectively, showing the closer to red, the higher the relative similarity to lung cancer cell exosomes. - Referring to (A) and (B) of
FIG. 9 , since large quantities of exosomes derived from normal cells are present in the lung cancer patient's blood, a majority of exosome data is found different from lung cancer cell exosomes, but compared to the normal persons, lung cancer patient exosome data shows that a higher percentage of exosomes is similar to lung cancer cell exosomes. - (C) of
FIG. 9 statistically shows the mean of 50 exosome data acquired from one person. A higher similarity to lung cancer cell exosomes is found in lung cancer patients than normal persons, and a p-value is 0.000036, and thus it can be seen that a statistical difference is very significant. -
FIG. 10 is a diagram showing a receiver operating characteristic (ROC) curve for a method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure. - The ROC curve is shown to validate the effect of the deep learning model according to the above-described embodiment of the present disclosure. As the area under the ROC curve, i.e., area under curve (AUC) is closer to 1, utility is higher, and the deep learning model according to an embodiment of the present disclosure shows a value of 0.89.
- The following Table 2 shows AUC value comparison results of the deep learning model according to an embodiment of the present disclosure and other research findings.
-
TABLE 2 Author Method AUC A. Hulbert, et. Al. PCR based DNA methylation detection 0.77-0.89 M. Aslam, et. Al. RCR based ctDNA detection 0.81 Y. Ni, et. Al. ELISA based apelin detection 0.81 W. Wang, et. Al. Circulating miRNA detection 0.71-0.87 Present disclosure — 0.89 - Referring to Table 2, according to an embodiment of the present disclosure, it can be seen that it has the top utility among liquid biopsy based lung cancer diagnosis methods. As a result, according to an embodiment of the present disclosure, it is possible to diagnose with a very high level of accuracy.
-
FIG. 11 is a diagram showing sensitivity and specificity by the method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure. - When a diagnostic cutoff point (cutoff line) is set to 0.85 based on the value derived from the ROC curve, sensitivity and specificity are about 89% and 88% respectively. That is, it can be seen that it is possible to clearly distinguish normal persons and lung cancer patients according to an embodiment of the present disclosure.
-
FIG. 12 is a diagram showing a result of staging lung cancer using the method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to an embodiment of the present disclosure, and in the graph, the stage is higher as it goes rightward. - Referring to
FIG. 12 , it can be seen that as the cancer stage is higher, the similarity to cancer cell exosomes is higher, and from this, it can be seen that the above-described embodiment of the present disclosure may be used to stage cancers. For example, it may be used in a manner of setting a similarity reference value by grade for each cancer stage and providing cancer stage information according to a range corresponding to the similarity derived by the process described above with reference toFIG. 1 . -
FIG. 13 is a block diagram of a system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to another embodiment of the present disclosure. - Referring to
FIG. 13 , thesystem 1000 for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to another embodiment of the present disclosure may include a trainingdata input unit 1100, a patientdata input unit 1200, adeep learning model 1300 and asimilarity analysis unit 1400. - The training
data input unit 1100 may receive input of exosome SERS signal data used to train thedeep learning model 1300. - According to an embodiment, the training
data input unit 1100 may receive input of SERS signals of exosomes derived from each of cultured cells, i.e., normal cells and cancer cells. - The patient
data input unit 1200 may receive input of blood exosome SERS signals of a patient for providing cancer diagnosis information. - The
deep learning model 1300 may first perform learning using the normal cell and cancer cell exosome SERS signals inputted through the trainingdata input unit 1100. - Additionally, after the
deep learning model 1300 is trained through the cultured cell exosome SERS signals, thedeep learning model 1300 may analyze the blood exosome SERS signals inputted through the patientdata input unit 1200. - The
similarity analysis unit 1400 may analyze a similarity between the blood exosome data and cultured cell exosome data analyzed through thedeep learning model 1300. - Here, the detailed configuration and/or operation method of each of the above-described elements is the same as the above description, and its repetitive description is omitted herein.
- The system for providing cancer diagnosis information shown in
FIG. 13 may be implemented in a computing device capable of deep learning computation. - The present disclosure is not limited to the above-described embodiments and the accompanying drawings. It is obvious to those having ordinary skill in the technical field pertaining to the present disclosure that substitution, modification and change may be made to the elements according to the present disclosure without departing from the technical spirit of the present disclosure.
Claims (7)
1. A method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome, comprising:
measuring a cultured cell exosome Surface Enhanced Raman Spectroscopy (SERS) signal;
training a deep learning model through the cultured cell exosome SERS signal;
measuring a blood exosome SERS signal;
analyzing the blood exosome SERS signal through the deep learning model trained with the cultured cell exosome SERS signal; and
analyzing a similarity between blood exosome data and cultured cell exosome data analyzed through the deep learning model.
2. The method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to claim 1 , wherein the cultured cell exosome SERS signal includes a normal cell exosome SERS signal and a cancer cell exosome SERS signal derived from cultured normal cell and cancer cell respectively.
3. The method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to claim 2 , wherein the deep learning model includes a first convolutional layer where data is inputted, a second convolutional layer, a third convolutional layer, a fourth convolutional layer, a fifth convolutional layer and a fully connected layer (FCL) connected to the first convolutional layer in a sequential order,
each of the second convolutional layer, the third convolutional layer, the fourth convolutional layer, the fifth convolutional layer and the fully connected layer includes a plurality of convolutional layers, and
the second to fifth convolutional layers add skip connection every two convolutional layers.
4. The method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to claim 3 , wherein the fully connected layer includes a first fully connected layer having 4096 values, and a second fully connected layer connected to an output end of the first fully connected layer and having 10 values.
5. The method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to claim 4 , wherein analyzing the similarity comprises calculating the similarity between the blood exosome data and the cancer cell exosome data by comparing a value of the second fully connected layer derived by analyzing the blood exosome SERS signal with a value of the second fully connected layer derived by analyzing the cultured cell exosome SERS signal.
6. The method for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome according to claim 5 , wherein analyzing the similarity comprises calculating a relative similarity of the blood exosome data to the cancer cell exosome data according to the following equation:
Relative similarity to cancer exosome=D normal /D cancer
Relative similarity to cancer exosome=D normal /D cancer
where the Dnormal is a Mahalanobis distance of the blood exosome data from the normal cell exosome data, and Dcancer is a Mahalanobis distance of the blood exosome data from the cancer cell exosome data.
7. A system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome, comprising:
a training data input unit to receive input of a cultured cell exosome Surface Enhanced Raman Spectroscopy (SERS) signal used to train a deep learning model;
a patient data input unit to receive input of a patient's blood exosome SERS signal data;
the deep learning model trained using the cultured cell exosome SERS signal inputted through the training data input unit to analyze the blood exosome SERS signal data inputted through the patient data input unit; and
a similarity analysis unit to analyze a similarity between blood exosome data and cultured cell exosome data analyzed through the deep learning model.
Applications Claiming Priority (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR1020190024870A KR102291105B1 (en) | 2019-03-04 | 2019-03-04 | Method and System of Providing Cancer Diagnosis Information Using Artificial Intelligence-Based Liquid Biopsy by Exosomes |
KR10-2019-0024870 | 2019-03-04 | ||
PCT/KR2020/000568 WO2020180003A1 (en) | 2019-03-04 | 2020-01-13 | Artificial intelligence-based method and system for provision of information on cancer diagnosis by using exosome-based liquid biopsy |
Publications (1)
Publication Number | Publication Date |
---|---|
US20220022816A1 true US20220022816A1 (en) | 2022-01-27 |
Family
ID=72337260
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US17/311,718 Pending US20220022816A1 (en) | 2019-03-04 | 2020-01-13 | Method and system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome |
Country Status (5)
Country | Link |
---|---|
US (1) | US20220022816A1 (en) |
EP (1) | EP3936031A4 (en) |
KR (1) | KR102291105B1 (en) |
CN (1) | CN113194820A (en) |
WO (1) | WO2020180003A1 (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CA3100035A1 (en) | 2018-08-27 | 2020-03-05 | Regeneron Pharmaceuticals, Inc. | Use of raman spectroscopy in downstream purification |
WO2022109240A1 (en) * | 2020-11-20 | 2022-05-27 | Sophos Limited | Classifier generator |
KR20220074088A (en) * | 2020-11-27 | 2022-06-03 | 주식회사 지씨지놈 | Method for diagnosing and predicting cancer type based on artificial intelligence |
KR20220160807A (en) * | 2021-05-28 | 2022-12-06 | 주식회사 지씨지놈 | Method for diagnosing and predicting cancer type based on cell-free nucleic acid and image analysis technology |
KR20220160806A (en) * | 2021-05-28 | 2022-12-06 | 주식회사 지씨지놈 | Method for diagnosing and predicting cancer type using fragment end motif frequency and size of cell-free nucleic acid |
KR102572517B1 (en) * | 2021-08-18 | 2023-08-31 | 주식회사 아폴론 | Method and apparatus for processing raman data of eosinophil based on artificial intelligence |
WO2023042646A1 (en) * | 2021-09-17 | 2023-03-23 | シンクサイト株式会社 | Classification model generation method, particle determination method, computer program, and information processing device |
WO2023042647A1 (en) * | 2021-09-17 | 2023-03-23 | シンクサイト株式会社 | Classification model generation method, particle classification method, computer program, and information processing device |
CN117916817A (en) * | 2022-08-18 | 2024-04-19 | 艾索波特株式会社 | Multi-tumor simultaneous diagnosis system and method based on artificial intelligence and using exosome SERS signals |
KR20240060943A (en) * | 2022-10-31 | 2024-05-08 | 주식회사 엑소퍼트 | System of Providing Artificial Intelligence Based mental illness Diagnosis Using Exosome SERS signals And Method Thereof |
CN117034110A (en) * | 2023-08-08 | 2023-11-10 | 海南省细胞医疗科技有限公司 | Stem cell exosome detection method based on deep learning |
Family Cites Families (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006271385A (en) * | 2001-03-14 | 2006-10-12 | Dna Chip Research Inc | Method of predicting colon cancer |
US20040009540A1 (en) * | 2001-04-27 | 2004-01-15 | Genoptix, Inc | Detection and evaluation of cancer cells using optophoretic analysis |
US7459696B2 (en) * | 2003-04-18 | 2008-12-02 | Schomacker Kevin T | Methods and apparatus for calibrating spectral data |
US20060281068A1 (en) * | 2005-06-09 | 2006-12-14 | Chemimage Corp. | Cytological methods for detecting a disease condition such as malignancy by Raman spectroscopic imaging |
KR100700913B1 (en) * | 2004-10-20 | 2007-03-28 | 고려대학교 산학협력단 | Method for reducing auto-fluorescence signals in confocal Raman microscopy |
US8034613B2 (en) * | 2005-06-01 | 2011-10-11 | Wisconsin Alumni Research Foundation | Multipotent lymphohematopoietic progenitor cells |
US7672702B2 (en) * | 2007-03-13 | 2010-03-02 | Samsung Electronics Co., Ltd. | Noninvasive in vivo measuring system and noninvasive in vivo measuring method by correcting influence of Hemoglobin |
EP2183382A1 (en) * | 2007-08-08 | 2010-05-12 | Chemimage Corporation | Raman difference spectra based disease classification |
EP2818549A4 (en) * | 2012-02-23 | 2015-09-02 | Sumitomo Bakelite Co | Method for classifying test body fluid sample |
US9942367B2 (en) * | 2015-10-13 | 2018-04-10 | Samsung Electronics Co., Ltd. | Electronic device and method for controlling the electronic device thereof |
US10746730B2 (en) * | 2015-12-18 | 2020-08-18 | Universiteit Gent | Method and system for characterizing extracellular vesicles |
US10376600B2 (en) * | 2016-03-03 | 2019-08-13 | Gholam A. Peyman | Early disease detection and therapy |
CN107303174A (en) * | 2016-04-21 | 2017-10-31 | 西派特(北京)科技有限公司 | A kind of internet+spectrum clinical tumor medical diagnostic method |
US11080591B2 (en) * | 2016-09-06 | 2021-08-03 | Deepmind Technologies Limited | Processing sequences using convolutional neural networks |
CN107271672B (en) * | 2017-06-21 | 2019-01-01 | 首都医科大学附属北京友谊医院 | Application of the excretion body p-ERK in preparation diagnosis of colorectal carcinoma product |
KR20180010280A (en) | 2018-01-16 | 2018-01-30 | 고려대학교 산학협력단 | Classification of Raman signals of exosomes by pattern recognition and the method of diagnosis of cells |
CN108537775A (en) * | 2018-03-02 | 2018-09-14 | 浙江工业大学 | A kind of cancer cell tracking based on deep learning detection |
KR101925603B1 (en) * | 2018-06-26 | 2019-02-26 | 주식회사 뷰노 | Method for faciliating to read pathology image and apparatus using the same |
-
2019
- 2019-03-04 KR KR1020190024870A patent/KR102291105B1/en active IP Right Grant
-
2020
- 2020-01-13 CN CN202080007009.9A patent/CN113194820A/en active Pending
- 2020-01-13 US US17/311,718 patent/US20220022816A1/en active Pending
- 2020-01-13 WO PCT/KR2020/000568 patent/WO2020180003A1/en unknown
- 2020-01-13 EP EP20767392.2A patent/EP3936031A4/en active Pending
Also Published As
Publication number | Publication date |
---|---|
WO2020180003A1 (en) | 2020-09-10 |
CN113194820A (en) | 2021-07-30 |
KR20200108938A (en) | 2020-09-22 |
KR102291105B1 (en) | 2021-08-23 |
EP3936031A1 (en) | 2022-01-12 |
EP3936031A4 (en) | 2022-12-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20220022816A1 (en) | Method and system for providing cancer diagnosis information using artificial intelligence-based liquid biopsy of exosome | |
Oh et al. | Identifying schizophrenia using structural MRI with a deep learning algorithm | |
Ortiz et al. | Exploratory graphical models of functional and structural connectivity patterns for Alzheimer's Disease diagnosis | |
Boschloo et al. | The Mood Disorder Questionnaire (MDQ) for detecting (hypo) manic episodes: its validity and impact of recall bias | |
CN110241205A (en) | A kind of schizophrenia biomarker combinations and its application and screening based on intestinal flora | |
Nakhleh et al. | Distinguishing idiopathic Parkinson's disease from other parkinsonian syndromes by breath test | |
WO2021108043A1 (en) | Assessing treatment response with estimated number of tumor cells | |
Romanò et al. | Fourier Transform Infrared Spectroscopy as a useful tool for the automated classification of cancer cell-derived exosomes obtained under different culture conditions | |
Roman et al. | Raman spectral signatures of urinary extracellular vesicles from diabetic patients and hyperglycemic endothelial cells as potential biomarkers in diabetes | |
Nabizadeh et al. | Artificial intelligence in the diagnosis of multiple sclerosis: A systematic review | |
CN111812078A (en) | Artificial intelligence assisted early diagnosis method for prostate tumor based on surface enhanced Raman spectroscopy | |
JP5847221B2 (en) | How to detect Alzheimer's disease risk | |
Erhardt et al. | Inflammatory biomarkers aid in diagnosis of dementia | |
EP3475705B1 (en) | Analysis of bodily fluids using infrared spectroscopy for the diagnosis and/or prognosis of cancer | |
JP6827464B2 (en) | Medical image | |
Du et al. | Diagnoses in multiple types of cancer based on serum Raman spectroscopy combined with a convolutional neural network: gastric cancer, colon cancer, rectal cancer, lung cancer | |
Zhang et al. | Clinical applicable AI system based on deep learning algorithm for differentiation of pulmonary infectious disease | |
US20220386952A1 (en) | Method to determine cognitive impairment | |
Wang et al. | Preliminary study on early detection technology of lung cancer based on surface-enhanced Raman spectroscopy | |
Zerweck et al. | An objective, automated and robust scoring using fluorescence optical imaging to evaluate changes in micro-vascularisation indicating early arthritis | |
CN115038967A (en) | Auxiliary method for early diagnosis of pancreatic cancer | |
CN111965238A (en) | Products, uses and methods for non-small cell lung cancer-related screening and assessment | |
CN117789972B (en) | Construction method of breast cancer recurrence prediction model and prediction system thereof | |
CN104237526B (en) | A kind of system detecting Alzheimer disease risk | |
Urbanavičiūtė et al. | Advanced image analysis-based evaluation of protein antibody microarray chemiluminescence signal improves glioma type identification by blood serum proteins concentrations |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: EXOPERT CORPORATION, KOREA, REPUBLIC OF Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:CHOI, YEON HO;SHIN, HYUN KU;PARK, JAE NA;AND OTHERS;REEL/FRAME:056464/0027 Effective date: 20210531 |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |